skip to navigation
skip to content

Kohlrabi 1.0.0

An asyncio-based task queue.

Kohlrabi is a asynchronous task queue for Python applications. It runs on top of a Redis server, and fits in with any existing application that runs on Python 3.3 or higher. It allows easy converting of parts of applications to asyncio-compatible parts, without affecting any normal blocking code.


Kohlrabi is available on PyPI to install easily:

pip install kohlrabi

Or, you can install it from GitHub directly:

pip install


Kohlrabi comes in two parts - the client and the server.

A single object is shared by both sides, and should be defined in your main file.

kh = kohlrabi.Kohlrabi()

When running the server, this object must be specified in a specific format on the command line:

kohlrabi-server yourapp.mainfile:kh
The first part before the : represents the import path of the module; how you would load it if you were to import it. The second part represents the Kohlrabi object created previously.
This will then load the Kohlrabi server, and load the tasks on the server-side.

In your app code, using Kohlrabi is incredibly simple.

Creating tasks

To create a task, simply decorate a function with Kohlrabi task decorator.

def hello():
    print("Hello, world!")

Then, inside your main method (or __name__ check), call the task as if it was a function.

if __name__ == "__main__":

If you check the console of the server process, it will have printed Hello, world!.

More advanced tasks

An example of a more advanced task would be an addition task.

def add(a, b):
    return a + b

Inside your main method, call the add task with the parameters chosen:

fut = add(1, 2)

This returns a ClientTaskResult object, which you can use to get the result of the task.

print("Added together 1 and 2 to get:", fut.result)

Note that ClientTaskResult.result is blocking, and will wait until the task has finished to get the result of the task. If you wish to only wait a certain amount of time, use the ClientTaskResult.result_with_timeout method.

print("Added together 1 and 2 to get:", fut.result_with_timeout(1))

Chaining tasks

If you wish to chain tasks together, use the yield from keywords. On the server side, a task is just a wrapped coroutine, meaning you can use it as if it was a coroutine.

def add_two(a):
    return a + 2

def get_four():
    four = yield from add_two(2)
    return four
File Type Py Version Uploaded on Size
Kohlrabi-1.0.0-py3-none-any.whl (md5, pgp) Python Wheel 3.5 2016-04-01 8KB
Kohlrabi-1.0.0.linux-x86_64.tar.gz (md5, pgp)
built for Linux-4.5.0-1-ARCH-x86_64-with-glibc2.3.4
"dumb" binary any 2016-04-01 11KB
Kohlrabi-1.0.0.tar.gz (md5, pgp) Source 2016-04-01 6KB